nkaje commented on pull request #162: URL: https://github.com/apache/mynewt-newtmgr/pull/162#issuecomment-636865498
> I made some (minor) comments. > > Overall I think this works, but I do think there will be some "weirdness" in how the code decides when to re-transmit a request. I think the only reliable way to determine if a retransmit is necessary is to match up sequence numbers between requests and responses (I briefly discussed this last time). > > The way it is implemented now (if I'm not misunderstanding it), the firmware has to report a bad offset 10 (11?) times before the code retransmits the missing chunk. Logically, the firmware should only need to report the issue once. If any given request yields a response with an unexpected offset, we immediately know there is a problem and that we need to retransmit. With this PR's implementation, we have to perform several useless transmits (useless because the firmware will immediately discard them) before we notice the problem. > > I feel like I have been enough of a pain in the neck with this PR :), so it is totally up to you whether this is a problem that needs to be fixed. I changed the watermark to -1. Perhaps it was only my gut feeling that dictated the original watermark. I tested with -1 and seems to work fine. ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org
